Abstract
This paper analyzes the relationship between theory and methods in that body of research that has come to be called "the new social history." First, an argument is made that the quantitative analysis of individual-level data that characterizes much of what is new in the recent sociohistorical literature is not only a methodology for conducting empirical research but also embodies an implicit theory of social organization that is both quantitative and individual. The power of such a theory to explain social change is limited, however, not only by technical problems associated with measurement and sampling error in historical research but also by the absence of an explicit causal component in the model. An alternative perspective, based on Marx's materialist theory of history, and which, it is argued, overcomes these limitations, is then presented. Empirical findings from research on the family in nineteenth century Los Angeles and Duncan's discussion of structural equation models, are used to illustrate how different theories of social change might be empirically tested.